Efficacy and safety of kinesiology tape for hemiplegic shoulder pain: A systematic review and meta-analysis of randomized controlled trials

2.1Search strategy

Electronic databases (Web of Science, MEDLINE, Embase, Cochrane Library, Wiley, Springer, Sciences Direct, Karger, PEDro, Scopus and CNKI, VIP, Wan Fang) were searched for articles published prior to December 13, 2020. The search terms were based on the PICOS principle (in Web of Science): (“Stroke” OR “shoulder pain” OR “hemiplegic shoulder pain” OR “shoulder subluxation” OR “shoulder-hand syndrome” OR “hand-shoulder syndrome” OR “hemiplegic hand”) AND (“Kinesio Taping” OR “kinesio tex tape” OR “Kinesiology tape” OR “Kinesio Tape” OR “Kinesiology Taping” OR “kinesio tex taping” OR “kinesiological taping” OR “Functional Fascial Taping” OR “Shoulder taping” OR “strapping”) AND (“placebo control” OR “bank control” OR “rehabilitation” OR “conventional treatment”) AND (“randomized controlled trials” OR “RCT” OR “randomized trials” OR “controlled trials”) in the title or abstract. The keywords were modified for the other databases (Supplementary Table 1). In addition, the reference lists of the retrieved articles were manually searched and reviewed. If discrepancies occurred, a consensus was reached through consultation. The searches were limited to studies published in English or Chinese.

2.2Selection criteria

Study eligibility was determined independently by two authors (LLX and ZY) according to the following criteria. (1) For the participants, all of the participants had been diagnosed with ischemic or hemorrhagic stroke by computerized tomography or magnetic resonance imaging and were clinically confirmed to have HSP. (2) For the interventions and comparisons, the KT technique was used in the treatment groups, and a sham KT or blank control group was included. (3) For the outcomes, the primary outcome in the study was pain, as assessed by the visual analog scale (VAS) and the numerical pain rating scale (NPRS). The secondary outcomes included the range of motion (ROM), acromion humeral distance (AHD, distance from the lower margin of the acromion to the head of the humerus), Fugl-Meyer assessment for upper extremities (FMA-UE) score, individual activity (IA, as assessed by the Barthel index (BI), modified Barthel index (MBI), Shoulder Pain and Disability Index (SPADI), and Action Research Arm Test (ARAT)), quality of life (QOL) and adverse events (AE). (4) For the study design, all the only randomized controlled trials (RCTs) that were in English or Chinese and aimed to investigate the efficacy and safety of KT were included. The exclusion criteria were as follows: (1) for the patients, the trial included patients with other conditions leading to hemiplegia (e.g., head injury, shoulder subluxation); (2) for the interventions, the trial administered complex treatment without specifying the sole effects of KT; (3) for the outcomes, the required data for this study were unavailable; and (4) for the design, the study was not an RCT. Disagreements were resolved by two authors (LLX and ZY) through discussion and consultation when necessary. A flowchart of the search process for this review was created in accordance with the 2009 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [17].

2.3Data extraction

Information on the following items was extracted from the included studies: first author, publication year, study period, age, sex, stroke duration, KT and control treatment protocols and measurement data. Based on the International Classification of Functioning, Disability and Health model (ICF) methods, data for the following four domains were extracted for this review: structure and function (pain, ROM, AHD, FMA-UE score), individual activities (MBI, SPADI, ARAT), social participation (QOL) and AEs. If the data sets overlapped or were duplicated, the articles with more information were retained. Disagreements were settled by discussion and consensus between the two authors (LLX and ZY).

2.4Quality assessment

The methodological quality of all eligible RCTs was assessed independently by the two reviewers (LLX and ZY) using the Physiotherapy Evidence Database Scale (PEDro) [18]. The PEDro scale has been shown to have good interpretability and internal and external validity and contains 11 dichotomous (yes or no) items on the eligibility criteria, random allocation, concealment of allocation, group similarity at baseline, blinding of the therapists, blinding of the assessors, availability of key outcome measures of more than 85% of the subjects, intention-to-treat analysis, between-group statistical comparisons, and point measures and measures of variability. One point is awarded when a criterion is satisfied, and the maximum score is 10 points. Studies with scores of 9–10 points were considered to have ‘excellent’ quality, those with scores of 6–8 were of ‘good’ quality, those with scores of 4–5 were of ‘fair’ quality, and those with scores of less than 4 were of ‘poor’ quality. We considered RCTs with a score of ⩾ 6 in this review, which typically represents a high-quality study [19]. The included RCTs were identified in the PEDro database (www.pedro.org.au) first and then manually scored if they were not found in the database. Any discrepancies in the ratings were resolved through discussion between two reviewers (LLX and ZY).

2.5Statistical analysis

The effect size of the continuous data was expressed as the standard mean difference (SMD) and 95% confidence interval (CI). The mean difference (MD) and pooled standard deviation (SD) were calculated based on the methods [20]. MD = (mean posttreatment - mean pretreatment), pooled SD =SC2,SC2=[(n1-1)⁢S12+(n2-1)⁢S22]/(n1+n2-2), (n1= sample size pretreatment; n2= sample size posttreatment; S1= SD pretreatment; S2= SD posttreatment). The I2 statistic was used to assess the level of statistical heterogeneity among the included studies; P< 0.10 and I2> 50% represented substantial heterogeneity [21]. The type of heterogeneity that existed was explained by using the following statistical methods: (1) sensitivity analysis; (2) subgroup analysis and meta-regression; and (3) a random-effects model using the DerSimonian-Laird method [22]. Publication bias was assessed with (1) funnel plots, (2) Begg’s test/Egger’s test, (3) a trim-and-fill adjusted analysis, and (4) the fail-safe number (Nfs) when necessary. The statistical analysis was executed by using a special edition of statistics and data science software (Stata SE, V.12.0; Stata Corp LP, College Station, TX, USA).

Figure 1.

3.1Study selection and characteristics

A total of 272 relevant references were initially retrieved. After the titles, abstracts and full texts were screened carefully, 14 studies [23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36] with 679 patients were ultimately included in the meta-analysis. A flowchart of the search process is shown in Fig. 1. The included trials were published from 2013 to 2020. Therapeutic KT was used in the experimental groups, and either sham KT or conventional rehabilitation methods such as exercise therapy and electrotherapy (blank control group) were performed in the control groups.

Among the included RCTs, the KT protocols differed substantially from each other. The tape remained in place from 15 minutes to 3 days, and the treatment duration ranged from 3 days to 8 weeks. The tape (5 cm width) was cut into a Y shape, X shape, or W shape or was not cut (I shape) and stretched with a tension of 10%–100% over the following four muscles: the supraspinatus, deltoid, brachii and trapezius. The sham tapes were applied at the same place with no tension. The conventional rehabilitation methods included physical therapy (passive and active assistance and active exercises), occupational therapy and electrotherapy. The outcomes included (1) pain (11 studies); (2) ROM (7 studies); (3) AHD (7 studies); (4) the FMA-UE score (7 studies); (5) IA (5 studies); (6) QOL (1 study); and (7) AE (0 studies). The characteristics of the included studies are detailed in Table 1. Based on the PEDro scale of bias tool, 12 studies had “good” methodological quality (6–8 scores). The quality assessment results are summarized in Table 2.

4.1Effect size of pain

Pain data assessed with VAS and NPRS were merged due to the high correlation [37]. The pooled effect size of pain (11 RCTs [24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34], 570 patients) showed significant heterogeneity (I=2 81.3%, P< 0.001), which disappeared after excluding 1 study [29] (SMD -6.50, 95% CI -8.32 to -4.69) via sensitivity analysis (I=2 48.5%, P= 0.042). Then, a fixed effect model was used, and the results showed that KT relieved pain in patients with HSP compared to the sham KT or blank control subgroup (SMD -0.92, 95% CI -1.10 to -0.74, P< 0.001, 10 RCTs, 539 patients), the forest plot is shown in Fig. 2A, and detailed information is available in Supplementary Table 2. The subgroup analysis showed that KT therapy was effective in reducing pain compared to sham KT or blank control, and there were no differences between the two subgroups (Sham KT: SMD -0.82, 95% CI -1.06 to -0.59, P< 0.001, 5 RCTs, 306 patients; Blank control: SMD -1.05, 95% CI -1.33 to -0.77, P< 0.001, 5 RCTs, 233 patients, respectively). A further publication bias analysis was performed. The funnel plot was visually symmetrical (Fig. 3A) and suggested that no obvious publication bias existed, which was also verified by Begg’s and Egger’s tests (P= 0.474, P= 0.689, respectively, detailed in Supplementary Table 3).

Figure 2.

Figure 3.

4.2Effect size of ROM

The effect size of ROM (7 studies, 320 patients) [23, 24, 25, 27, 28, 29, 31] was pooled with a random model and yielded positive results with significant heterogeneity (I=2 90.0%, P< 0.001, SMD 2.27, 95% CI 1.23 to 3.32, P< 0.001). The forest plot of ROM suggested that KT improved the range of motion of the shoulder joint for HSP (Fig. 2B, more information in Supplementary Table 2). Subgroup analyses showed that there was no difference between the KT and blank control groups (SMD 4.01, 95% CI -2.01 to 10.04, P= 0.192; I=2 96.9%; P< 0.001; 2 studies, 61 patients) in improving shoulder ROM in patients with HSP. In the analysis of publication bias analysis, the funnel plot was visually symmetrical (Fig. 3B) and confirmed by Begg’s and Egger’s tests (P= 1.000 and P= 0.996, respectively, Supplementary Table 3).

4.3Effect size of AHD

The pooled effect size of AHD (7 RCTs, 246 patients, with fixed model) [24, 25, 27, 28, 30, 35, 36] showed that the KT reduced the distance from the acromion to the humeral head and prevented subluxation of the shoulder joint in patients with HSP without significant heterogeneity (I=2 23.1%, P= 0.253, SMD -0.62, 95% CI -0.88 to -0.37, P< 0.001). The subgroup analysis suggested that KT reduced the AHD compared with sham KT or blank control (SMD -0.66, 95% CI -1.03 to -0.29, P< 0.001, 4 RCTs, 124 patients; SMD -0.59, 95% CI -0.95 to -0.22, P= 0.002, 3 RCTs, 122 patients;, respectively). The forest plot of AHD is shown in Fig. 2C, and more information is detailed in Supplementary Table 2. In turn, publication bias analyses were conducted, and the funnel plot was first visually symmetrical (Fig. 3C). Then, Begg’s and Egger’s tests were performed, and no publication bias was observed (P= 0.548 and P= 0.149, respectively, Supplementary Table 3).

4.4Effect size of FMA-UE

The FMA-UE data extracted from 7 studies [27, 28, 32, 33, 34, 35, 36], (383 patients) were pooled with a random effect model and showed positive results with considerable heterogeneity (I=2 85.2%, P< 0.001). Two studies [32, 34] were excluded via sensitivity analysis (I=2 41.9%, P= 0.142). The pooled effect size of the remaining and excluded studies both showed that KT improved upper extremity motor function (SMD 0.66, 95% CI 0.41 to 0.91, P< 0.001, 5 RCTs, 263 patients; 2 RCTs, 120 patients, respectively), which are detailed in forest plots (Fig. 2D). Then, subgroup analysis was conducted, and the results showed that KT significantly improved upper extremity motor function compared with both sham KT and blank control (SMD 0.48, 95% CI 0.04 to 0.93, P< 0.05, 2 RCTs, 84 patients; SMD 0.74, 95% CI 0.43 to 1.04, P< 0.001, 3 RCTs, 179 patients; Supplementary Table 2). Finally, publication bias analyses were performed, and the funnel plot was symmetrical in visual inspection and was verified by Begg’s and Egger’s tests (P= 1.000 and P= 0.946, respectively, Supplementary Table 3).

4.5Effect size of IA and QOL

The IA data were synthesized from the BI, MBI, and SPADI data (5 RCTs [25, 26, 27, 32, 34], 232 patients). A higher score indicates more severe pain and disability on the SPADI scale, while it indicates more independence on the BI and MBI. Hence, the MD of the SPADI score was subtracted by the maximum (100) and reversed in the same direction as BI and MBI. The pooled effect size (including subgroup analysis) with a random effect model showed no improvement in individual daily living activity (SMD 0.42, 95% CI -0.22 to 1.07, P= 0.199; I=2 77.8%, P= 0.001; 5 RCTs, 199 patients; Fig. 2E, detailed in Supplementary Table 2). Then, the publication bias analyses were not conducted. The meta-analysis of QOL was not performed because it was only performed in 1 study [26] with negative results (SMD 0.26, 95% CI -0.34 to 0.85, P= 0.393; SMD 0.31, 95% CI -0.29 to 0.90, P= 0.311). More information is detailed in Supplementary Table 2.

4.6Adverse events

In all included studies, it was reported that no adverse events related to KT therapy occurred. Therefore, it is safe to use KT for the treatment of HSP.

5.1Limitations

There were several limitations in the present study. First, the number of RCTs and the sample size were still small. Second, all the studies included in this meta-analysis reported only short-term effects (3–4 weeks). Third, this study focused on patients with chronic convalescent stroke within 1 to 6 months after stroke onset. Therefore, no conclusions can be drawn for patients with acute or subacute stroke. Fourth, there was diversity of and variation in the KT protocols. The selected taping muscles, taping time and frequency, treatment duration and taping shape and tension showed significant heterogeneity in the included studies. Therefore, more rigorously designed RCTs with larger sample sizes and with longer follow-up evaluations are needed in the future. Standardized protocols of KT therapy should also be recommended.

5.2Implication

To the best of our knowledge, this is the first meta-analysis to evaluate the effectiveness and safety of KT therapy versus sham KT for treating HSP. In the present study, after rigorous meta-analysis, it was concluded that KT did have significant effects on reducing pain, increasing ROM, and shortening the subluxation distance of the shoulder joint but not on improving daily living activity and quality of life for hemiplegic shoulder pain without side effects. Interestingly, we also found that four muscles (supraspinatus, deltoid, brachii and trapezius) were most commonly used to treat HSP. The findings were similar to those of other published reviews.